UAS Pointclouds vs. Lidar Pointclouds for structual analysis of forests

Introduction

Use of LiDAR in forestry and nature conservation

The use of Light detection and ranging (LiDAR) pointclouds are well established in forestry, agriculture and forest research. Spatially extensive estimations of vegetation related structural parameters are mostly realised with the calculation of LiDAR indices in a regular grid. Common applications on a forest stand scale are the retrieval of canopy cover and heights (J. Lee et al. 2018; Alexander et al. 2014), stand density (C.-C. Lee and Wang 2018) and the estimation of leaf area (Kamoske et al. 2019). The structural information and its heterogeneity in a landscape serve as indicators for biodiversity (Hilmers et al. 2018) or species occurrence (Carrasco et al. 2019; Melin et al. 2016; Froidevaux et al. 2016). Most of these Lidar indices are strongly correlated (Shi et al. 2018).

Lidar change detection in trees (Duncanson and Dubayah 2018)

Tree height from field measurements and Lidar (Jurjević et al. 2020)

Problems with LiDAR

Despite their relevance in forest applications, Lidar data has some major drawbacks, mainly in their cost and accessibility. Lidar sensors and data acquisition are expensive and often distributed commercially. Data provided by governmental institutions are for the most part still irregularly available and not publicly distributed. Further, the temporal resolution of the data is low (by law every 3 years in Germany) making them not suitable for monitoring or applications which require different seasonal conditions.

Comparison of leaf-on leaf-off Lidar (Davison, Donoghue, and Galiatsatos 2020) might be a problem, since we have no control with official lidar data

Seasonality of Lidar: max and mean height independent of seasonality, sd and skew of height more dependent on phenology

Tree height works good in UAS (Fawcett 2019)

Monitoring of Canopy height of crops with multitemporal UAS based CHMs, DEM was build manually with known ground point interpolation (Grüner 2019)

also comparing well to TLS based pointclouds when evaluating plant height in agriculture (Malambo 2018)

Understory trees are a big problem and are not detected in individual tree segmentations (< 35% Goldberg 2018) hence we dont do it here

Use of UAS digital aerial photogrammetry

With the recent development of unmanned aerial systems (UAS) and photogrammetric techniques like structure from motion (SfM), an alternative to Lidar pointclouds are available. Quick data access in moderately large areas makes UAS data promising for the monitoring of agricultural or natural systems (Manfreda et al. 2018). Depending on flight conditions, these pointclouds could be acquired on a near daily basis. Especially in forest environments, research can benefit from vegetation structural information retrieved from UAS data.

mix of pointclouds and images to classify individual trees (Xu et al. 2020)

However, if trees are detected, the height measurement is well established and consistent across multiple flight dates (e.g. Krause et al. 2019)

Previous comparisons of UAS and Lidar

comparison of lidar and uav CHM (Michez et al. 2020)

Comparison of Lidar and UAS based pointclouds for individual tree height (Ganz, Käber, and Adler 2019) revealed very good compariability. Quality of UAS based studies is highly depended on the accuracy of data acuisition and ulitmately proper georeferencing. Dealed with that in Ludwig 2020!

Comparisons revealed good potential of UAS pointclouds as a substitude for lidar when estimating common forest attributes (e.g. Ullah et al. 2019; Cao et al. 2019) and to a lesser extent biomass estimations in the tropics (Ota et al. 2015)

DEM clearly is the week point of photogrammetry (Ota et al. 2015) When comparing ALS and UAS, using a ALS derived DTM is common in order to normalize the Pointclouds (e.g. Ullah et al. 2019)

UAS pointclouds do not have return values which many Lidar indices depend on. Every point is a first return so we cannot get below a developed canopy. These return values however are crucial for LiDAR point classification (e.g. differentiate between ground and non ground point).

EBV framework with 3D information: height, cover and structural complexity Heterogenous data sources: requires the comparability of Lidar and photogrammetrically recieved pointclouds (Valbuena et al. 2020)

The quality and viability of UAS pointclouds have to be assessed in terms of comparability to Lidar pointclouds (since Lidar structural analysis is the standard in many studies)

Previous work of multitemporal UAS

Multitemporal UAV for monitoring coral reefs (Fallati et al. 2020)

Multitemporal UAS can benefit monitoring, e.g. tree growth rates (Guerra-Hernández et al. 2017) or crops (Moeckel et al. 2018)

Multitemporal UAS orthoimages can enhance classification of vegetation types slightly (van Iersel et al. 2018), makes use of plant phenology, most imprtant were july and september because there, green vegetation were at the maximum

Biomass of single trees (in a park) much better under leaf off conditions (Ye, van Leeuwen, and Nyktas 2019)

If the positional accuracy of the individual photogrammetric pointclouds is high enough (previously shown in Ludwig et al. (2020), it is a resonable assumption to combine pointclouds from different phenological stages in order to get a full 3D model of the forest.

Since photogrammetically received pointclouds only capture the surface and do not penetrate the forest canopy like Lidar pointclouds, different phenological stages should capture different vertical layers of the forest canopy.

What we do

This study demonstates the usability of multitemporal pointclouds derived from digital aerial photogrammetry for forest structural analysis. Commonly applied forest structural indicators will be compared between DAP and LiDAR pointclouds for different spatial scales and different phenological stages of a deciduous forest. In addition we propose the combination of multiple DAP pointclouds as a way to improve their information value and better comparability to LiDAR data. All derived pointcloud indicators will also be related to commonly used forest structural indicators from field surveys of trees.

Hypothesis 1: When compared to LiDAR pointclouds, the quality of structural indices from DAP pointclouds depend on the phenological stages of a deciduous forest.

Hypothesis 2: Multitemporal DAP pointclouds are superior than monotemporal pointclouds and are suitable to complement LiDAR derived pointclouds for forest structural analysis.

Methods

Study Area

The study area is a 200 x 150 m part of a mixed deciduous forest near Marburg (Germany). The area consists of a mix of oaks () and beeches () and represent a typical environment in a managed forest. The elevation ranges from XXXm to XXXm a.s.l. Stem positions of 500 trees were acquired by using a differential GPS (Zenith 35 Pro, GeoMax Widnau Switzerland) with a positioning accuracy of 0.05 m.\

Datasets

ID	Type	Sensor	Acquisition Date Y-M-D	Description
1	LiDAR	Riegel LMS-Q780	2018-04-06	leaf off
2	DAP	GoPro Hero7	2020-04-18	early leaf development in lower canopy
3	DAP	GoPro Hero7	2020-04-22	advanced leaf development in lower canopy
4	DAP	GoPro Hero7	2020-09-15	full canopy
5	DAP	GoPro Hero7	2020-10-28	full canopy / beginning of coloring
6	DAP	GoPro Hero7	2020-10-31	early leaf fall
7	DAP	GoPro Hero7	2020-11-03	advanced leaf fall
8	DAP	GoPro Hero7	2020-11-12	advanced leaf fall
9	DAP	GoPro Hero7	2020-12-10	leaf off

The LiDAR pointcloud (provided by the Hessian Agency for Nature Conservation, Environment and Geology - HLNUG) was in early spring 2018 under leaf off conditions. The used Sensor was a Riegl LMS-Q780 with a maximum space between points of 0.85 m and the positional accuracy from a Novatel OEM4 GNSS system was 0.3 m horizontally and 0.15 m vertically.

The DAP pointclouds were acquired with a 3DR Solo Quadrocopter (3D Robotics, Inc., Berkeley CA, USA) and a GoPro Hero 7 camera (GoPro Inc., San Mateo CA, USA).

The individual images were processed using the photogrammetric software Metashape (Agisoft LLC, St. Petersburg, Russia) using the workflow described in (Ludwig et al. 2020).

Pointcloud preprocessing and combination

Common lidar indices

Canopy cover

Usually canopy cover is derived from LiDAR as follows (over 10 different studies cited in Bakx et al. (2019) Supplementary Material 3)

\[ \frac{N_{p > x}}{N_{t}} * 100 \]

with percentage of returns (Np > x) above x meter above ground level at the raster resolution. Nt is the total number of returns. Bakx et al. (2019) also mentiones Farrell et al. 2013 in which a tow part procedure is described: First cover is estimated from aerial photographs, then it is corrected by excluding areas with low canopy height derived from LiDAR. UAS based pointclouds might very suitable for this approach, since the pointcloud and the aerial image are received in the same workflow.

Canopy height

Maximum canopy height (z_max)

Highest LiDAR return in a raster cell (over 10 different studies cited in Bakx et al. (2019) Supplementary Material 3)

\[ Z_{max} \]

Mean canopy height 95% (z_mean95)

Mean height of the returns in the 95 percentile (Z95). N95 is the number of returns in the 95 percentile

\[ Z_{mean95} = \frac{\Sigma(Z_{95})}{N_{95}} \]

Mean canopy height (z_mean_csm)

Mean height of the canopy surface model (CSM) in the grid cell (first return of the LiDAR). For Gap correction only points above a certain threshold are used (over 10 different studies cited in Bakx et al. (2019) Supplementary Material 3)

Horizontal canopy variability (index_name_sd)

Usually the standard deviation of the canopy cover or canopy height in larger raster cell (e.g. 10 m - reasonable to get to the sentinel scale!)

Vertical canopy variability

Coefficient of variation of canopy height (CV)

Ratio between mean canopy height (Zmean) and standard deviation (Zsd) of canopy height (5 different studies cited in Bakx et al. (2019))

\[ CV = \frac{Z_{mean}}{Z_{sd}} \]

Standard deviation of canopy height (z_sd_csm)

Standard deviation of first returns in a raster cell (over 10 different studies cited in Bakx et al. (2019) Supplementary Material 3)

Results

Direct comparison of Lidar and DAP

Comparison of the vertical point distribution of LiDAR and single date DAP

Comparison of the vertical point distribution of LiDAR and combined DAP in spring and fall

Common Indices for different resolutions

Voxel mean height at different resolutions

Horizontal heterogenity

Multitemporal pointclouds

Date Summary

all these plots use a 2m resolution for the indice calculation

Discussion

Outlook: posibility of individual tree segmentation and then relate to stand variable (Sackov 2019) or even tree health (Belmonte 2018)

Text fragments

The main challange for further usage of Lidar data in a forest environment is the detection of individual trees. This enables the estimation of tree related parameters such as diameter at breast height, timber volume or crown related metrics (van Leeuwen and Nieuwenhuis 2010). Forest structure then can be described as the sum of the structure of individual trees (e.g. their height and biomass Ferraz et al. 2016) and the species composition (REF). This could give new insights into ecosystem functioning, since many processes and species distributions depend on functions provided by trees or their related microhabitats (REF). Further, monitoring of individual tree health and drought could be applied in forestry.

References

Alexander, Cici, Peder Klith Bøcher, Lars Arge, and Jens-Christian Svenning. 2014. “Regional-Scale Mapping of Tree Cover, Height and Main Phenological Tree Types Using Airborne Laser Scanning Data.” Remote Sensing of Environment 147 (May): 156–72. https://doi.org/10.1016/j.rse.2014.02.013.

Bakx, Tristan R. M., Zsófia Koma, Arie C. Seijmonsbergen, and W. Daniel Kissling. 2019. “Use and Categorization of Light Detection and Ranging Vegetation Metrics in Avian Diversity and Species Distribution Research.” Edited by Damaris Zurell. Diversity and Distributions 25 (7): 1045–59. https://doi.org/10.1111/ddi.12915.

Cao, Lin, Hao Liu, Xiaoyao Fu, Zhengnan Zhang, Xin Shen, and Honghua Ruan. 2019. “Comparison of UAV LiDAR and Digital Aerial Photogrammetry Point Clouds for Estimating Forest Structural Attributes in Subtropical Planted Forests.” Forests 10 (2): 145. https://doi.org/10.3390/f10020145.

Carrasco, Luis, Xingli Giam, Monica Papeş, and Kimberly Sheldon. 2019. “Metrics of Lidar-Derived 3d Vegetation Structure Reveal Contrasting Effects of Horizontal and Vertical Forest Heterogeneity on Bird Species Richness.” Remote Sensing 11 (7): 743. https://doi.org/10.3390/rs11070743.

Davison, Sophie, Daniel N. M. Donoghue, and Nikolaos Galiatsatos. 2020. “The Effect of Leaf-on and Leaf-Off Forest Canopy Conditions on LiDAR Derived Estimations of Forest Structural Diversity.” International Journal of Applied Earth Observation and Geoinformation 92 (October): 102160. https://doi.org/10.1016/j.jag.2020.102160.

Duncanson, Laura, and Ralph Dubayah. 2018. “Monitoring Individual Tree-Based Change with Airborne Lidar.” Ecology and Evolution 8 (10): 5079–89. https://doi.org/10.1002/ece3.4075.

Fallati, Luca, Luca Saponari, Alessandra Savini, Fabio Marchese, Cesare Corselli, and Paolo Galli. 2020. “Multi-Temporal UAV Data and Object-Based Image Analysis (OBIA) for Estimation of Substrate Changes in a Post-Bleaching Scenario on a Maldivian Reef.” Remote Sensing 12 (13): 2093. https://doi.org/10.3390/rs12132093.

Ferraz, António, Sassan Saatchi, Clément Mallet, and Victoria Meyer. 2016. “Lidar Detection of Individual Tree Size in Tropical Forests.” Remote Sensing of Environment 183 (September): 318–33. https://doi.org/10.1016/j.rse.2016.05.028.

Froidevaux, Jérémy S. P., Florian Zellweger, Kurt Bollmann, Gareth Jones, and Martin K. Obrist. 2016. “From Field Surveys to LiDAR: Shining a Light on How Bats Respond to Forest Structure.” Remote Sensing of Environment 175 (March): 242–50. https://doi.org/10.1016/j.rse.2015.12.038.

Ganz, Selina, Yannek Käber, and Petra Adler. 2019. “Measuring Tree Height with Remote SensingA Comparison of Photogrammetric and LiDAR Data with Different Field Measurements.” Forests 10 (8): 694. https://doi.org/10.3390/f10080694.

Guerra-Hernández, Juan, Eduardo González-Ferreiro, Vicente Monleón, Sonia Faias, Margarida Tomé, and Ramón Díaz-Varela. 2017. “Use of Multi-Temporal UAV-Derived Imagery for Estimating Individual Tree Growth in Pinus Pinea Stands.” Forests 8 (8): 300. https://doi.org/10.3390/f8080300.

Hilmers, Torben, Nicolas Friess, Claus Bässler, Marco Heurich, Roland Brandl, Hans Pretzsch, Rupert Seidl, and Jörg Müller. 2018. “Biodiversity Along Temperate Forest Succession.” Edited by Nathalie Butt. Journal of Applied Ecology 55 (6): 2756–66. https://doi.org/10.1111/1365-2664.13238.

Jurjević, Luka, Xinlian Liang, Mateo Gašparović, and Ivan Balenović. 2020. “Is Field-Measured Tree Height as Reliable as Believed Part II, A Comparison Study of Tree Height Estimates from Conventional Field Measurement and Low-Cost Close-Range Remote Sensing in a Deciduous Forest.” ISPRS Journal of Photogrammetry and Remote Sensing 169 (November): 227–41. https://doi.org/10.1016/j.isprsjprs.2020.09.014.

Kamoske, Aaron G., Kyla M. Dahlin, Scott C. Stark, and Shawn P. Serbin. 2019. “Leaf Area Density from Airborne LiDAR: Comparing Sensors and Resolutions in a Temperate Broadleaf Forest Ecosystem.” Forest Ecology and Management 433 (February): 364–75. https://doi.org/10.1016/j.foreco.2018.11.017.

Krause, Stuart, Tanja G. M. Sanders, Jan-Peter Mund, and Klaus Greve. 2019. “UAV-Based Photogrammetric Tree Height Measurement for Intensive Forest Monitoring.” Remote Sensing 11 (7): 758. https://doi.org/10.3390/rs11070758.

Lee, Chung-Cheng, and Chi-Kuei Wang. 2018. “Estimating Stand Density in a Tropical Broadleaf Forest Using Airborne LiDAR Data.” Forests 9 (8): 475. https://doi.org/10.3390/f9080475.

Lee, Junghee, Jungho Im, Kyungmin Kim, and Lindi Quackenbush. 2018. “Machine Learning Approaches for Estimating Forest Stand Height Using Plot-Based Observations and Airborne LiDAR Data.” Forests 9 (5): 268. https://doi.org/10.3390/f9050268.

Ludwig, Marvin, Christian M. Runge, Nicolas Friess, Tiziana L. Koch, Sebastian Richter, Simon Seyfried, Luise Wraase, et al. 2020. “Quality Assessment of Photogrammetric MethodsA Workflow for Reproducible UAS Orthomosaics.” Remote Sensing 12 (22): 3831. https://doi.org/10.3390/rs12223831.

Manfreda, Salvatore, Matthew F McCabe, Pauline E Miller, Richard Lucas, Victor Pajuelo Madrigal, Giorgos Mallinis, Eyal Ben Dor, et al. 2018. “On the Use of Unmanned Aerial Systems for Environmental Monitoring,” 28.

Melin, M., J. Matala, L. Mehtätalo, J. Pusenius, and P. Packalen. 2016. “Ecological Dimensions of Airborne Laser Scanning Analyzing the Role of Forest Structure in Moose Habitat Use Within a Year.” Remote Sensing of Environment 173 (February): 238–47. https://doi.org/10.1016/j.rse.2015.07.025.

Michez, Adrien, Leo Huylenbroeck, Corentin Bolyn, Nicolas Latte, Sébastien Bauwens, and Philippe Lejeune. 2020. “Can Regional Aerial Images from Orthophoto Surveys Produce High Quality Photogrammetric Canopy Height Model? A Single Tree Approach in Western Europe.” International Journal of Applied Earth Observation and Geoinformation 92 (October): 102190. https://doi.org/10.1016/j.jag.2020.102190.

Moeckel, Thomas, Supriya Dayananda, Rama Nidamanuri, Sunil Nautiyal, Nagaraju Hanumaiah, Andreas Buerkert, and Michael Wachendorf. 2018. “Estimation of Vegetable Crop Parameter by Multi-Temporal UAV-Borne Images.” Remote Sensing 10 (5): 805. https://doi.org/10.3390/rs10050805.

Ota, Tetsuji, Miyuki Ogawa, Katsuto Shimizu, Tsuyoshi Kajisa, Nobuya Mizoue, Shigejiro Yoshida, Gen Takao, et al. 2015. “Aboveground Biomass Estimation Using Structure from Motion Approach with Aerial Photographs in a Seasonal Tropical Forest.” Forests 6 (12): 3882–98. https://doi.org/10.3390/f6113882.

Shi, Yifang, Tiejun Wang, Andrew K. Skidmore, and Marco Heurich. 2018. “Important LiDAR Metrics for Discriminating Forest Tree Species in Central Europe.” ISPRS Journal of Photogrammetry and Remote Sensing 137 (March): 163–74. https://doi.org/10.1016/j.isprsjprs.2018.02.002.

Ullah, Sami, Matthias Dees, Pawan Datta, Petra Adler, Mathias Schardt, and Barbara Koch. 2019. “Potential of Modern Photogrammetry Versus Airborne Laser Scanning for Estimating Forest Variables in a Mountain Environment.” Remote Sensing 11 (6): 661. https://doi.org/10.3390/rs11060661.

Valbuena, R., B. O’Connor, F. Zellweger, W. Simonson, P. Vihervaara, M. Maltamo, C. A. Silva, et al. 2020. “Standardizing Ecosystem Morphological Traits from 3d Information Sources.” Trends in Ecology & Evolution 35 (8): 656–67. https://doi.org/10.1016/j.tree.2020.03.006.

van Iersel, Wimala, Menno Straatsma, Hans Middelkoop, and Elisabeth Addink. 2018. “Multitemporal Classification of River Floodplain Vegetation Using Time Series of UAV Images.” Remote Sensing 10 (7): 1144. https://doi.org/10.3390/rs10071144.

van Leeuwen, Martin, and Maarten Nieuwenhuis. 2010. “Retrieval of Forest Structural Parameters Using LiDAR Remote Sensing.” European Journal of Forest Research 129 (4): 749–70. https://doi.org/10.1007/s10342-010-0381-4.

Xu, Zhong, Xin Shen, Lin Cao, Nicholas C. Coops, Tristan R. H. Goodbody, Tai Zhong, Weidong Zhao, et al. 2020. “Tree Species Classification Using UAS-Based Digital Aerial Photogrammetry Point Clouds and Multispectral Imageries in Subtropical Natural Forests.” International Journal of Applied Earth Observation and Geoinformation 92 (October): 102173. https://doi.org/10.1016/j.jag.2020.102173.

Ye, Ning, Louise van Leeuwen, and Panagiotis Nyktas. 2019. “Analysing the Potential of UAV Point Cloud as Input in Quantitative Structure Modelling for Assessment of Woody Biomass of Single Trees.” International Journal of Applied Earth Observation and Geoinformation 81 (September): 47–57. https://doi.org/10.1016/j.jag.2019.05.010.